The complex mix of proteins found in serum ("the serum proteome") may provide valuable information about ongoing pathophysiologic processes and the presence of subclinical disease. Analysis of serum proteomic patterns measured by mass spectrometry, in particular, has shown promise in detecting such pathophysiologic disturbances as ovarian and prostate cancer. Since atherosclerosis is primarily an intravascular process, we hypothesize that it too may be amenable to serum proteomic pattern analysis. The primary aim of this proposal, therefore, is to test the hypothesis that serum proteomic patterns will reflect the presence or absence of subclinical atherosclerosis. In order to identify serum proteomic patterns associated with atherosclerosis, we plan to conduct a case-control study of 200 patients with, and 200 without angiographic evidence of atherosclerosis whose serum was previously collected for the UCSF Genomic Resource in Arteriosclerosis. Blindly labeled sera will be analyzed with new ultra-sensitive electrospray ionization mass spectrometry. The resulting spectra will be processed and then analyzed using a variety of classification and discrimination techniques to find patterns that identify patients with and without atherosclerosis. We will use the first 100 cases and 100 controls to develop a decision rule, then use the second 100 cases and 100 controls to validate the rule. The ability of this rule to discriminate between patients with and without atherosclerosis will be tested against rules utilizing standard coronary heart disease risk factor data. Peaks in the spectrum that are particularly strong discriminators will be used to identify the specific serum proteins that characterize atherosclerosis. This research is designed to yield insight into the pathogenesis of coronary artery disease, a serum test for atherosclerosis, and a start towards unraveling the interplay between genes, proteins and phenotypic coronary heart disease, the leading cause of death in the US.